Low Vitamin Intake Is Associated with Risk of Frailty in Older Adults

Age and Ageing 2018; 0: 1–8 © The Author(s) 2018. Published by Oxford University Press on behalf of the British Geriatrics Society. doi: 10.1093/ageing/afy105 All rights reserved. For permissions, please email: [email protected]

Low vitamin intake is associated with risk of frailty in older adults

1 2,3 2,3,4 2,3 TERESA BALBOA-CASTILLO ,ELLEN A. STRUIJK ,ESTHER LOPEZ-GARCIA ,JOSÉ R. BANEGAS , 2,3,4 2,3,4,5 FERNANDO RODRÍGUEZ-ARTALEJO ,PILAR GUALLAR-CASTILLON

1Department of Public Health—EPICYN Research Center, School of Medicine, Universidad de La Frontera, Temuco, Chile 2Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid-IdiPaz, Madrid, Spain 3CIBERESP (CIBER of Epidemiology and Public Health), Madrid, Spain 4IMDEA-Food Institute, CEI UAM+CSIC, Madrid, Spain 5Johns Hopkins Bloomerg School of Public Health, Baltimore, USA Address correspondence to: T. Balboa-Castillo. Tel: +56 45 2734172; Fax: +56 45 2592139. Email: [email protected]

Abstract

Background: the association between vitamin intake and frailty has hardly been studied. The objective was to assess the association of dietary vitamin intake with incident frailty in older adults from Spain. Methods: data came from a cohort of 1,643 community-dwelling individuals aged ≥65, recruited in 2008–10 and followed up prospectively throughout 2012. At baseline, 10 vitamins were assessed (vitamin A, thiamine, riboflavin, niacin, vitamins B6, B12, C, D, E and folates) using a validated face-to-face diet history. Incident frailty was identified using Fried’sdefinition as having ≥3 of the following five criteria: unintentional weight loss of ≥4.5 kg, exhaustion, weakness, slow walking speed and low physical activity. Nonadherence to the recommended dietary allowances (RDA) was considered when the intake of a vitamin was below the recommendation. Analyses were performed with logistic regression and adjusted for main confounders. Results: during a 3.5-year follow-up, 89 (5.4%) participants developed frailty. The odds ratios (95% confidence interval) of frailty for those in the lowest versus the highest tertile of vitamin intake were 2.80 (1.38–5.67), P-trend: 0.004, for vitamin B6; 1.65 (0.93–2.95), P-trend: 0.007, for vitamin C; 1.93 (0.99–3.83), P-trend: 0.06, for vitamin E and 2.34 (1.21–4.52), P-trend: 0.01, for folates. Nonadherence to the RDAs of vitamins was related to frailty for thiamine odds ratio (OR): 2.09 (1.03–4.23); niacin OR: 2.80 (1.46–5.38) and vitamin B6; 2.23 (1.30–3.83). When considering tertiles of RDAs for the 10 vitamins those who met <5 RDAs had a higher risk of frailty, OR: 2.84 (1.34–6.03); P-trend: <0.001, compared to those who met >7. Conclusion: a lower intake of vitamins B6, C, E and folates was associated with a higher risk of frailty. Not meeting RDAs for vitamins was also strongly associated.

Keywords: frailty, vitamin intake, older people, recommended dietary allowances, diet quality

Introduction adults in clinical settings as well as for research purposes. Identifying frail older adults is of relevance because frailty is Frailty is a geriatric syndrome that has sarcopenia as its patho- potentially reversible after being treated with physical activity physiological basis. Age-related musculoskeletal changes and and an appropriate diet [6]. anorexia of ageing are also considered two main pathways for There is evidence that poor nutrition is related to frailty this syndrome. To be frail implies some degree of functional [1, 7, 8], so, to some extent, frailty can also be considered damage, and frail older adults experience a decrease in func- as a nutrition-related condition [9]. Some physiological tional reserve and a lack of adequate response even to minor changes occur in the gastrointestinal system with ageing stressors (e.g. a cold, diarrhoea or dehydration). As a conse- (such as alterations in taste and smell, reduction of gastric quence, frail subjects have increased vulnerability to developing motility and changes in gastrointestinal hormones). All adverse health outcomes, including falls, disabilities and mor- these changes can modify dietary preferences, dietary intake tality [1–4]. Fried et al. [5] provided an operational deﬁnition and impair absorption of macro and micronutrients, leading to identify frailty that can be used to recognise frail older to malnutrition and vitamin deﬁciencies. In addition, as we

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018 T. Balboa-Castillo et al.

get older, poor adherence to a healthy diet can arise due to were asked to indicate how frequently each specificfoodand social reasons such as lack of access to fresh food or diffi- beverage was habitually consumed in the preceding year. culties in handling and cooking of food. However, the lon- This instrument records the consumption of 880 foods, with gitudinal association of vitamin intake with incident frailty a set of coloured photographs to help in the quantification has hardly being studied among the elderly. of portions. Data on food consumption were converted into A cross-sectional analysis of the InCHIANTI Study assessed daily intake energy, macronutrients and micronutrients using the association of vitamin intake with prevalent frailty. This ana- standard food composition tables [16–19]. The intake of 10 a lysis included individuals aged 65 and older and found that a priori relevant vitamins below were assessed: vitamin A, thia- low intake of vitamins D, E, C and folates was independently mine, riboflavin, niacin, vitamin B6, vitamin B12, vitamin C, related to prevalent frailty [10]. Also, one longitudinal study, the vitamin D, vitamin E and folates. Participants were cate- Women’s Health and Aging Study I (WHAS I), assessed the gorisedintosex-specific tertiles for vitamin intake. relationship of some serum vitamins with incident frailty among Nonadherence to the recommended dietary allowances disabled women; a low concentration of carotenoids and vitamin (RDA) was considered when the intake of a vitamin was E was associated with increased risk of frailty [11]. below the recommendation. To assess nonadherence to the Therefore, vitamin intake and its association with frailty RDAs, the Spanish recommendations for older population have not been prospectively and systematically studied in [16] were used: vitamin A μg/day: ≥1,000 in men or ≥800 older adults, neither in men nor in women. Consequently, in women; thiamine mg/day: ≥1inmenor≥0.8 in women; this study assessed the prospective association between riboflavin mg/day: ≥1.4 in men or ≥1.1 in women; niacin both, dietary vitamin intake with incident frailty in older mg/day: ≥16 in men or ≥12 in women; vitamin B6 mg/day: adults from Spain. ≥1.8 in men or ≥1.6 in women; vitamin B12 μg/day: ≥2in men or women; vitamin C mg/day: ≥60 in men or women; vitamin D ≥ percentile 75 in men or women; vitamin E mg/ Methods day: ≥12 in men or women and folates μg/day: ≥400 in Study design and participants men or women. Those who met the corresponding RDA for the Spanish older population were used as reference. Data were taken from the Seniors-ENRICA cohort, its Finally, by summing up the number of RDAs met for – fl methods have been previously reported [12 14]. Brie y, the each participant, we obtained a score ranging from 0 to 10 – cohort was established in 2008 10 with 2,519 individuals that was classified in tertiles according to the number of fi selected by strati ed random sampling from the population recommendations met: <5, 5 to 7 and >7. Those who met aged 60 and older in Spain. At baseline, data were collected >7 recommendations were used as reference. in three stages. Through a phone interview, we obtained information on sociodemographic variables, lifestyle, health Frailty status and morbidity, followed by a home visit to collect fi fi blood and urine samples; finally, a second home visit to car- We used a minor modi cation of the operational de nition of frailty developed by Fried and colleagues in the cardiovas- ry out a physical examination and to collect dietary infor- fi fi mation. Participants were followed up throughout 2012, cular health study [5]. Speci cally, frailty was de ned as having three or more of the following five criteria: (i) weight when a second wave of data collection was performed to fi ≥ update information on frailty. loss, de ned as unintentional loss of 4.5 kg of body weight For the current analysis, we excluded 372 individuals with- in the preceding year; (ii) exhaustion was evaluated as a posi- out baseline information on diet or with implausibly high or tive answer to either of the following two questions taken – from the Centre for Epidemiologic Studies Depression Scale: low energy intake (outside the range of 800 5,000 kcal/d for ‘ ’ ‘ men or 500–4,000 kcal/d for women), plus 66 participants Do you feel that anything you did was a big effort? and Do you feel you could not keep on doing things?’ at least 3–4 who were taking vitamin supplements. We also excluded 359 fi individuals with prevalent frailty or with limitations in instru- days a week; (iii) weakness, de ned as the lowest quintile in mental activities of daily living measured with the Lawton and the cardiovascular health study of maximum strength of the Brody scale. Of the remaining 1,722 participants, we addition- dominant hand, adjusted for sex and body mass index ally excluded 79 with missing values for confounders. Thus, (BMI). Strength was measured using a Jamar dynamometer, the analyses were conducted with 1,643 individuals. and the highest value in two consecutive measurements was selected; (iv) slow walking speed, defined as the lowest Study participants gave written informed consent. The fi study was approved by the Clinical Research Ethics cohort-speci cquintileinourstudysampleforthethree- Committee of La Paz University Hospital in Madrid. metre walking speed test, adjusted for sex and height and (v) low physical activity, defined as walking ≤2.5 h/week in men and ≤2 h/week in women. Study variables Vitamin intake Other variables At baseline, information on food consumption was obtained At baseline, we obtained information on sociodemographic using a validated face-to-face dietary history [15]. Participants variables, anthropometrics, lifestyle and prevalent chronic

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018 Vitamin intake and frailty

diseases. Specifically, study participants reported their sex, there was also a linear trend for the risk of frailty across ter- age and educational level (primary, secondary and univer- tiles (P = 0.007) (Table 2). sity). Weight and height were measured under standardised The achievement of RDAs for vitamins fluctuated conditions. BMI was calculated as weight (kg) divided by between 22% for folates to 98% for vitamin B12. height in metres squared. Participants also gave information Nonadherence to the RDAs was related to frailty for thiamine on smoking status (never, former and current smoker). (OR, 2.09; 95% CI, 1.03–4.23), niacin (2.80; 1.46–5.38) and Sedentary behaviour was approximated by television view- vitamin B6 (2.23; 1.30–3.83) (Supplementary Table 1, available ing time (h/week). Total alcohol (g/day) and energy intake in Age and Ageing online). Furthermore, when comparing with (kcal/d) were estimated with standard food composition those who met RDAs for >7 vitamins, those who met <5 tables from Spain [16–19]. They also reported whether they vitamins had a higher risk of frailty (OR, 2.84; 95% [CI], 1.34 suffered from any of the following physician-diagnosed dis- to 6.03) (Supplementary Table 2, available in Age and Ageing eases: coronary heart disease, stroke, cancer, chronic pul- online). monary disease and type-2 diabetes mellitus. When the dose–response association was depicted, the risk of frailty was linear for vitamin B6, C and E. However, for folates, the risk of frailty initially decreases when folate Statistical analysis intake increases, down to a plateau around an intake of The associations between tertiles of vitamin intake and inci- 350 mg/day (Figure 1). dent frailty were summarised with ORs and their 95% confidence intervals (95% CI), obtained from multivariate logistic regression. The upper tertile of consumption was used as Discussion reference. Logistic models were adjusted for sex, age, educational level, BMI, smoking status, alcohol intake, television In the present analysis conducted with community-dwelling viewing time, total energy intake, coronary heart disease, older adults, both poor intake of vitamins B6, C, E and stroke, cancer, chronic pulmonary disease and type-2 diabetes folates and nonadherence to RDAs for thiamine, niacin and mellitus. Linear trend was assessed modelling the tertiles of vitamin B6 were independently associated with incident vitamin intake as a continuous variable. Additionally, when in frailty. Meeting a lower number of RDAs for vitamins was the logistic analysis, the P for trend was statistically signifi- also strongly associated with a higher risk of frailty. cant, we also evaluated the dose–response relationship In a cross-sectional analysis conducted with 802 indivi- between vitamin intake and frailty by fitting nonparametric duals aged 65 and over from the InCHIANTI study [10], regression curves (cubic splines) adjusted for confounders. low intake of vitamins D, E, C and folates was independ- Other logistic models were built to assess the association ently related to prevalent frailty. Our results are consistent between tertiles of the number of RDAs met at baseline and with this study except for the fact that we did not find a sig- incident frailty. nificant association for vitamin D. Moreover, our findings Statistical significance was set at two-tailed P < 0.05. are in line with other analysis from the InCHIANTI [20], Analyses were conducted using the SAS software, version showing that participants in the plasma highest vitamin E 9.4 (SAS Institute Inc.) and Stata/SE, version 11.1 tertile were less likely to be frail than those participants in (StataCorp, College Station, TX, USA). the lowest vitamin E tertile. Only the WHAS I [11] study has previously analysed the Results association of some serum micronutrients and vitamins with incident frailty. In this analysis, conducted with 766 Over a mean follow-up time of 3.5 years, 89 (5.4%) partici- moderately to severely disabled community-dwelling older pants developed frailty. Compared with nonfrail individuals women, the risk of frailty was higher in those with low those with frailty were more often women, and older, had a levels of carotenoids and vitamin E. We obtained similar lower educational level and a higher BMI, spent more time results for vitamin E, but not for vitamin B6 or folates. watching TV and showed a higher frequency of coronary Frail participants from the WHAS I study had higher serum heart disease, stroke and diabetes. Also, they had lower concentrations of vitamin B6 and folates than nonfrail par- intake of alcohol and of most vitamins (including vitamin ticipants, however, the difference was not statistically signifi- A, thiamine, niacin, vitamins B6, B12, D, E and folates) cant. Vitamin C was not measured in the WHAS I study. (Table 1). In our analyses, vitamin deficiencies were related to In fully adjusted analyses, the ORs (95% CI) of incident frailty, and the risk was higher in older people that met few- frailty for those in the lowest versus the upper tertile of er number of RDAs for vitamins. This is in line with some vitamin intake were 2.80 (1.38–5.67) P-trend: 0.004 for vita- previous studies reporting a relationship between better diet min B6, 1.93 (1.00–3.83) P-trend: 0.06 for vitamin E and quality and lower risk of frailty [21, 22]. Thus, high diet 2.34 (1.21–4.52) P-trend: 0.01 for folates. The association quality was inversely associated with prevalent and incident was not statistically significant for the rest of vitamins, frailty in older men [7], and also a diet that includes con- though the ORs were always above 1. For vitamin C intake, sumption of fruit, vegetables and dairy products has been

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018 T. Balboa-Castillo et al.

Table 1. Baseline characteristics of the Seniors-ENRICA cohort study according to incident frailty status (N = 1,646).

Total Not frail Frail (N = 1,646) (n = 1,557) (n = 89) ...... Characteristics Mean (SD) or n (%) Mean (SD) or n (%) Mean (SD) or n (%) P Men 825 (50.1) 797 (51.2) 28 (31.5) 0.003 Age, years 68.1 (5.9) 67.8 (5.8) 72.2 (6.9) <0.0001 Educational level Primary 863 (52.4) 793 (50.9) 70 (78.7) <0.0001 Secondary 426 (25.9) 417 (26.8) 9 (10.1) University 357 (21.7) 347 (22.3) 10 (11.2) Body mass index (kg/m2) 28.4 (4.2) 28.2 (4.1) 31 (5.2) <0.0001 Smoking status Never smoker 934 (56.7) 876 (56.3) 58 (65.2) 0.095 Former smoker 509 (30.9) 483 (31.0) 26 (29.2) Current smoker 203 (12.3) 198 (12.7) 5 (5.6) Television viewing time (h/week) 17.4 (10.8) 17.1 (1.06) 22.3 (13.3) <0.0001 Alcohol intake (g/day) 11.2 (18.4) 11.46 (18.6) 6.4 (14.6) 0.01 Total energy intake (kcal/day) 2,079 (745) 2,088 (750) 1,936 (629) 0.06 Prevalent diseases Coronary heart disease 22 (1.3) 18 (1.2) 4 (4.5) 0.007 Stroke 15 (0.9) 12 (0.8) 3 (3.4) 0.01 Cancer 27 (1.6) 25 (1.6) 2 (2.3) 0.64 Chronic obstructive pulmonary disease 115 (7.0) 105 (6.7) 10 (11.2) 0.11 Type-2 diabetes mellitus 235 (14.3) 213 (13.7) 22 (24.7) 0.004 Vitamin intake Vitamin A (μg/day) 869 (604.2) 874.7 (613.2) 775.5 (407.0) 0.13 Thiamine (mg/day) 1.4 (0.5) 1.4 (0.5) 1.3 (0.4) 0.04 Riboﬂavin (mg/day) 1.6 (0.5) 1.6 (0.5) 1.6 (0.6) 0.34 Niacin (mg/day) 21.1 (7.3) 21.2 (7.3) 19.3 (8.1) 0.02 Vitamin B6 (mg/day) 2.0 (0.6) 2.01 (0.6) 1.9 (0.6) 0.004 Vitamin B12 (μg/day) 6.4 (3.8) 6.4 (3.8) 5.3 (2.4) 0.006 Vitamin C (mg/day) 137.6 (70.7) 138.3 (70.7) 125.0 (70.0) 0.08 Vitamin D (μg/day) 3.4 (3.1) 3.5 (3.1) 2.8 (1.8) 0.03 Vitamin E (mg/day) 10.1 (5.1) 10.2 (5.1) 8.9 (3.8) 0.01 Folates (μg/day) 326.7 (112.6) 328.7 (112.5) 293.4 (111.1) 0.004

associated with a lower risk of developing frailty in the focused on vitamin D and suggested that the relationship short term [23, 24]. between vitamin D status and frailty is mediated by sarco- No data have been published on the association between penia [26]. Scott et al. [27] have shown that 25-hydroxy- meeting RDAs for vitamins and frailty. Our results suggest vitamin D increased muscle mass and strength over a that deficiencies of several B vitamins are associated with 2-year follow-up. Furthermore, vitamin D stimulates mus- frailty. Among them, deficiency for B6 might be considered cle recovery, which is associated with an increase in muscle the most important because it affected 30% of the sample. mass and strength in the elderly [28]. In cross-sectional Moreover, although deficiencies for vitamin E and folates studies, the association of serum 25-hydroxyvitamin D defi- did not reach statistical significance when RDA cutoff ciency and frailty has been frequently observed [29–31]; but points were considered, the intake of these vitamins is also in prospective studies, circulating 25-hydroxy-vitamin D of potential relevance because of their negative dose– levels was weakly associated with frailty risk [32, 33]. response association with frailty when were classified in ter- Although vitamin D intake does not take into account the tiles, and because the RDAs for those vitamins are not effect of sun exposure, our results are in line with a weak reached by the majority of the population; in our sample, association. less than 35% of the participants satisfied the RDAs for Saum et al. [34] have reported a major role of oxidative vitamins A, D, E and folates. Also, in the SENECA study, stress and inflammation in the development of frailty. So, a considerable proportion of older adults had a deficiency antioxidant vitamins such as vitamins A, C and E could in vitamin A, thiamine and riboflavin, despite an adequate play a role in frailty prevention. Also, Matteini et al. [35] nutritional status assessed by total energy intake and BMI found that vitamin B12 deficiency contributed to frailty in [25]. older women and they suggested that some genetic poly- Vitamins may play a role in the pathogenesis of frailty morphisms may lead to decreased vitamin B12 availability through several biological pathways. Most studies have and contribute to the risk of frailty. Concerning vitamin

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018 Vitamin intake and frailty

Table 2. Odds ratios (95% interval, 95% CI) for the association between vitamin intake (tertiles) at baseline and incident frailty (N = 1,646).

Vitamins intake Vitamin A Thiamine Riboﬂavin Niacin Vitamin B6 ...... N/frail OR (95% CI) N/frail OR (95% CI) N/frail OR (95% CI) N/frail OR (95% CI) N/frail OR (95% CI) Tertile 1 (upper) 549/26 1.00 (Ref.) 549/27 1.00 (Ref.) 549/24 1.00 (Ref.) 549/24 1.00 (Ref.) 549/16 1.00 (Ref.) Tertile 2 549/28 1.08 (0.62–1.90) 549/30 1.14 (0.63–2.07) 549/32 1.16 (0.66–2.07) 549/29 1.27 (0.71–2.28) 549/32 1.30 (0.75–2.25) Tertile 3 (lower) 548/35 1.17 (0.64–2.15) 548/32 1.30 (0.68–2.51) 548/33 1.37 (0.73–2.61) 548/36 1.42 (0.73–2.77) 548/41 2.80 (1.38–5.67)** P for trend 0.61 0.42 0.32 0.29 0.004

Vitamin B12 Vitamin C Vitamin D Vitamin E Folates ...... N/Frail OR, (95% CI) N/Frail OR, (95% CI) N/Frail OR, (95% CI) N/Frail OR, (95% CI) N/Frail OR, (95% CI) Tertile 1 (upper) 549/27 1.00 (Ref.) 549/23 1.00 (Ref.) 549/26 1.00 (Ref.) 549/21 1.00 (Ref.) 549/19 1.00 (Ref.) Tertile 2 549/26 1.57 (0.89–2.79) 549/27 1.40 (0.82–2.41) 549/29 1.07 (0.61–1.86) 549/31 1.31 (0.75–2.30) 549/28 1.51 (0.87–2.63) Tertile 3 (lower) 548/36 1.20 (0.66–2.17) 548/39 1.65 (0.93–2.95) 548/34 1.16 (0.66–2.05) 548/37 1.93 (0.99–3.83)* 548/42 2.34 (1.21–4.52)* P for trend 0.48 0.007 0.61 0.06 0.01

*P < 0.05; **P < 0.01. Model adjusted for sex, age (years), educational level (primary, secondary, university), body mass index (quartiles kg/m2), smoking status (never smoker, former smoker, current smoker), alcohol intake (g/day), television viewing time (quartiles weekly hours), total energy consumption (quartiles kcal/day), coronary heart disease, stroke, cancer, chronic obstructive pulmonary disease and type-2 diabetes mellitus. Tertiles cutoff points: Vitamin A μg/day: 645; 1,054 in men and 571; 977 in women. Thiamine mg/day: 1.2; 1.7 in men and 1.0; 1.5 in women. Riboﬂavin mg/day: 1.5; 2.0 in men and 1.3; 1.8 in women. Niacin mg/day: 19.6; 26.9 in men and 16.0; 22.9 in women. Vitamin B6 mg/day: 1.9; 2.6 in men and 1.6; 2.2 in women. Vitamin B12 μg/day: 4.9; 8.7 in men and 3.9; 6.9 in women. Vitamin C mg/day: 98.1; 177.0 in men and 99.9; 179.9 in women. Vitamin D μg/day: 2.1; 5.1 in men and 1.7; 4.2 in women. Vitamin E mg/day: 8.2; 13.2 in men and 6.9; 11.7 in women. Folates μg/day: 281.6; 415.9 in men and 254.0; 370.3 in women.

B12, our results show a nonlinear negative association vitamins in reducing mortality were much stronger in parti- with the risk of frailty without achieving statistical cipants with vitamin insufficiency [40]. Our findings, how- significance. ever, were obtained with vitamins through the diet, and In addition, all vitamins related to frailty in our analyses they do not endorse the use of pharmacological dose sup- are considered essential, and each one could play a patho- plements that should only be used in the case of important genic role in frailty affecting muscle formation or mainten- deficiencies. ance. As such, vitamin B6 is related to the metabolism of Finally, this study has several strengths such as being a amino acids; vitamin C is important in the formation of longitudinal study of noninstitutionalised Spanish older collagen; vitamin E is involved in the formation of muscles adults. In addition, food consumption, specifically vitamin and other tissues and folates are involved in cellular intake, was assessed with a validated computerised dietary division. history, that showed good correlation for nutrients and Our results are of public health relevance for several vitamins [15]. Also, many potential confounders were reasons. First, because the prevalence of frailty and its accounted for. Some limitations should also be acknowl- clinical importance are increasing even in developing edged. We did not measure serum vitamin levels, as in countries [36]; Second, because nonadherence to RDA many observational study, despite adjusting for many values for vitamins could warn us of the need for asses- potential confounders, some residual confounding may sing frailty status. Lastly, because an inadequate nutritional persist. In the same way, between-person variability in vita- intake is an important modifiable risk factor for frailty, it min intake cannot be ruled out as part of measurement could be palliated by achieving the RDAs for B vitamins errors. and increasing intake of some essential vitamins on a regu- In conclusion, in this prospective study, in older adults, lar dietary basis. poor vitamin intake for vitamin B6, C, E and folates and Also, it is worth noticing that some clinical trials have nonadherence to the RDAs were strongly and independently found that taking daily multivitamins did not reduce major associated with frailty in the elderly. Both situations could be cardiovascular events, cancer outcomes and cardiovascular easily identified in clinical settings through a nutritional mortality [37–39]. However, the expected effects from interview.

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018 T. Balboa-Castillo et al.

Frailty incidence Frailty incidence 10.0 20 10.0 20

4.0 4.0 15 15

2.0 2.0

10 10 1.0 1.0 Odds Ratio Odds Ratio Percentage Percentage

0.5 0.5 5 5

0.25 0.25

0 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 3 3.5 4 4.5 5 5.5 6 Logarithm of Vitamin B6 intake (mg/day) Logarithm of Vitamin C intake (mg/day)

Frailty incidence Frailty incidence 10.0 20 10.0 20

4.0 4.0 15 15

2.0 2.0

10 10 1.0 1.0 Odds Ratio Percentage Odds Ratio Percentage 0.5 0.5 5 5

0.25 0.25

0 0 0 0.5 1 1.5 2.0 2.5 3.0 3.5 4 4.5 5 5.5 6 6.5 7 7.5 Logarithm of Vitamin E intake (mg/day) Logarithm Folates intake (mg/day) Figure 1. Dose–response association between vitamin intake and frailty (N = 1,646). Y-axis represents the OR for incident frailty and X-axis the logarithm of vitamin intake. Dashed lines are 95% conﬁdence intervals. The bars represent the frequency of vitamin intake at baseline. Model adjusted for sex, age (years), educational level (primary, secondary, university), body mass index (quartiles kg/m2), smoking status (never smoker, former smoker, current smoker), alcohol intake (g/day), television viewing time (quartiles weekly hours), total energy consumption (quartiles kcal/day), coronary heart disease, stroke, cancer, chronic obstructive pulmonary disease and type-2 diabetes mellitus.

Authors’ contributions Key points TBC and PGC had full access to all of the data in the study • Poor vitamin intake and non-adherence to the RDAs for and take responsibility for the integrity of the data and the vitamins were strongly associated with frailty in the elderly. accuracy of the data analysis. Study concept and design: TBC • Low vitamin intake could be easily identified in clinical and PGC. Acquisition of data: PGC, FRA and ELG. Statistical settings through a nutritional interview to prevent frailty analysis: TBC and PGC. Interpretation of data: TBC and in elderly. PGC. Draughting of the manuscript: TBC and PGC. Critical • Deficiencies of several B vitamins are associated with revision of the manuscript for important intellectual content: frailty in elderly. EAS, ELG, JRB and FRA. Study supervision: PGC. All authors have read and approved the final manuscript. Supplementary data fl Supplementary data mentioned in the text are available to Con ict of interest subscribers in Age and Ageing online. None.

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018 Vitamin intake and frailty

Funding 14. Struijk EA, Guallar-Castillón P, Rodríguez-Artalejo F, López- García E. Mediterranean Dietary Patterns and Impaired Data collection was funded by the following grants: FIS PI16/ Physical Function in Older Adults. J Gerontol A Biol Sci 1512; PI16/1460; PI16/609; PI17/1709 (State Secretary of R Med Sci 2016; 73: 333–9. +D and FEDER/FSE) Spanish Ministry of Economy and 15. Guallar-Castillón P, Sagardui-Villamor J, Balboa-Castillo T Competitiveness DEP2013-47786-R grant FEDER/FSE, the et al. Validity and reproducibility of a Spanish dietary history. FRAILOMIC Initiative (EU FP7-HEALTH-2012-Proposal PLoS One 2014; 9: e86074. no. 305483-2), the ATHLOS project (EU H2020-Project ID: 16. Moreiras O, Carvajal A, Cabrera L et al. Tablas de composición de alimentos. (Ediciones Pirámide, 2007). 635316), the SALAMANDER project (PCIN-2016-145) and 17. Farrán A, Zamora R, Cervera P. Tablas de composición de alimen- the CIBERESP, Instituto de Salud Carlos III. Madrid, Spain. tos del CESNID. (Edicions Universitat de Barcelona, 2004). 18. Ortega RM, Requejo AM, Andrés P et al. Dietary intake and cognitive function in a group of elderly people. Am J Clin References Nutr 1997; 66: 803–9. 19. Mataix Verdú J, García Díz L, Mañas Almendros M, 1. Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty Martinez de Vitoria E, Llopis González J. Tabla de composición in elderly people. Lancet 2013; 381: 752–62. de alimentos. 4th ed. (Editorial Universidad de Granada., 2003). 2. Macklai NS, Spagnoli J, Junod J, Santos-Eggimann B. 20. Ble A, Cherubini A, Volpato S et al. Lower plasma vitamin E Prospective association of the SHARE-operationalized frailty levels are associated with the frailty syndrome: the phenotype with adverse health outcomes: evidence from 60+ InCHIANTI study. J Gerontol A Biol Sci Med Sci 2006; 61: community-dwelling Europeans living in 11 countries. BMC 278–83. Geriatr 2013; 13: 3. 21. Chan R, Leung J, Woo J. Dietary Patterns and Risk of Frailty 3. Vermeulen J, Neyens JCL, van Rossum E, Spreeuwenberg in Chinese Community-Dwelling Older People in Hong Kong: MD, de Witte LP. Predicting ADL disability in community- A Prospective Cohort Study. Nutrients 2015; 7: 7070–84. dwelling elderly people using physical frailty indicators: a sys- 22. León-Muñoz LM, Guallar-Castillón P, López-García E, tematic review. BMC Geriatr 2011; 11: 33. Rodríguez-Artalejo F. Mediterranean diet and risk of frailty 4. Woo J, Leung J, Morley JE. Comparison of frailty indicators in community-dwelling older adults. J Am Med Dir Assoc based on clinical phenotype and the multiple deficit approach 2014; 15: 899–903. in predicting mortality and physical limitation. J Am Geriatr 23. Garcia-Esquinas E, Rahi B, Peres K et al. Consumption of Soc 2012; 60: 1478–86. fruit and vegetables and risk of frailty: a dose-response ana- 5. Fried LP, Tangen CM, Walston J et al. Frailty in older adults: lysis of 3 prospective cohorts of community-dwelling older evidence for a phenotype. J Gerontol A Biol Sci Med Sci adults. Am J Clin Nutr 2016; 104: 132–42. 2001; 56: M146–56. 24. Lana A, Rodriguez-Artalejo F, Lopez-Garcia E. Dairy con- 6. Rolland Y, Dupuy C, Abellan van Kan G, Gillette S, Vellas sumption and risk of frailty in older adults: a prospective B. Treatment strategies for sarcopenia and frailty. Med Clin cohort study. J Am Geriatr Soc 2015; 63: 1852–60. North Am 2011; 95: 427–38. 25. Toffanello ED, Inelmen EM, Minicuci N et al. Ten-year 7. Shikany JM, Barrett-Connor E, Ensrud KE et al. trends in vitamin intake in free-living healthy elderly people: Macronutrients, diet quality, and frailty in older men. the risk of subclinical malnutrition. J Nutr Health Aging J Gerontol A Biol Sci Med Sci 2014; 69: 695–701. 2011; 15: 99–103. 8. Bollwein J, Volkert D, Diekmann R et al. Nutritional status 26. Bruyère O, Cavalier E, Buckinx F, Reginster J-Y. Relevance according to the mini nutritional assessment (MNA®) and of vitamin D in the pathogenesis and therapy of frailty. Curr frailty in community dwelling older persons: a close relation- Opin Clin Nutr Metab Care 2016; 20: 26–9. 10.1097/MCO. ship. J Nutr Health Aging 2013; 17: 351–6. 0000000000000334. 9. Cederholm T, Barazzoni R, Austin P et al. ESPEN guidelines 27. Scott D, Blizzard L, Fell J, Ding C, Winzenberg T, Jones G. on definitions and terminology of clinical nutrition. Clin Nutr A prospective study of the associations between 25-hydroxy- 2016; 36: 49–64. doi:10.1016/j.clnu.2016.09.004. vitamin D, sarcopenia progression and physical activity in 10. Bartali B, Frongillo EA, Bandinelli S et al. Low nutrient intake older adults. Clin Endocrinol (Oxf) 2010; 73: 581–7. is an essential component of frailty in older persons. 28. Domingues-Faria C, Vasson M-P, Goncalves-Mendes N, J Gerontol A Biol Sci Med Sci 2006; 61: 589–93. Boirie Y, Walrand S. Skeletal muscle regeneration and impact 11. Semba RD, Bartali B, Zhou J, Blaum C, Ko CW, Fried LP. of aging and nutrition. Ageing Res Rev 2016; 26: 22–36. Low serum micronutrient concentrations predict frailty 29. Hirani V, Naganathan V, Cumming RG et al. Associations among older women living in the community. J Gerontol A between frailty and serum 25-hydroxyvitamin D and 1,25- Biol Sci Med Sci 2006; 61: 594–9. dihydroxyvitamin D concentrations in older Australian men: 12. Rodríguez-Artalejo F, Graciani A, Guallar-Castillón P et al. the Concord Health and Ageing in Men Project. J Gerontol [Rationale and methods of the study on nutrition and cardio- A Biol Sci Med Sci 2013; 68: 1112–21. vascular risk in Spain (ENRICA)]. Rev Esp Cardiol 2011; 64: 30. Tajar A, Lee DM, Pye SR et al. The association of frailty with 876–82. serum 25-hydroxyvitamin D and parathyroid hormone levels 13. Struijk EA, Lana A, Guallar-Castillón P, Rodríguez-Artalejo in older European men. Age Ageing 2013; 42: 352–9. F, Lopez-Garcia E. Intake of B vitamins and impairment in 31. Wilhelm-Leen ER, Hall YN, Deboer IH, Chertow GM. physical function in older adults. Clin Nutr 2016; 37: 1271–8. Vitamin D deficiency and frailty in older Americans. J Intern doi:10.1016/j.clnu.2017.05.016. Med 2010; 268: 171–80.

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018 T. Balboa-Castillo et al.

32. Ensrud KE, Ewing SK, Fredman L et al. Circulating 25- 37. Sesso HD, Buring JE, Christen WG et al. Vitamins E and C hydroxyvitamin D levels and frailty status in older women. in the prevention of cardiovascular disease in men. JAMA J Clin Endocrinol Metab 2010; 95: 5266–73. 2008; 300: 2123–33. 33. Ensrud KE, Blackwell TL, Cauley JA et al. Circulating 25- 38. Sesso HD, Christen WG, Bubes V et al. Multivitamins in the preven- hydroxyvitamin D levels and frailty in older men: the osteo- tion of cardiovascular disease in men: the Physicians’ Health Study II porotic fractures in men study. J Am Geriatr Soc 2011; 59: randomized controlled trial . J Am Med Assoc 2013; 308: 1751–60. 101–6. 39. Schwingshackl L et al. Dietary supplements and risk of cause- 34. Saum K-U, Dieffenbach AK, Jansen EH et al. Association speciﬁc death, cardiovascular disease, and cancer: a systematic between oxidative stress and frailty in an elderly german review and meta-analysis of primary prevention trials. Adv Nutr population: results from the ESTHER cohort study. 2017; 8: 27–39. 10.3945/an.116.013516.Noncommunicable. Gerontology 2015; 61: 407–15. 40. Brenner H, Jansen L, Saum K-U, Holleczek B, Schöttker B. 35. Matteini AM, Walston JD, Fallin MD et al. Markers of B- Vitamin D supplementation trials aimed at reducing mortality vitamin deﬁciency and frailty in older women. J Nutr Health have much higher power when focusing on people with low Aging 2008; 12: 303–8. serum 25-hydroxyvitamin D concentrations. J Nutr 2017; 36. Nguyen TN, Cumming RG, Hilmer SN. A review of frailty in 147: 1325–33. developing countries. J Nutr Health Aging 2015; 19: 941–6. Received 15 January 2018; editorial decision 22 June 2018

Downloaded from https://academic.oup.com/ageing/advance-article-abstract/doi/10.1093/ageing/afy105/5058977 by UNIVERSITY OF CALIFORNIA, Berkeley user on 26 July 2018